Design, synthesis and biological evaluation of acridone analogues as novel STING receptor agonists

Article abstract of DOI:10.1016/j.bioorg.2019.103556

STING (Stimulator of Interferon Genes) has become a focal point in immunology research and a target in drug discovery. The discovery of a potent human-STING agonist is expected to revolutionize current anti-virus or cancer immunotherapy. Inspired by the structure and function of murine STING-specific agonists (DMXAA and CMA), we rationally designed and synthesized four series of novel compounds for the enhancement of human sensitivity. In the cell-based assay, we identified six compounds from all the synthetic small molecules: 2g, 9g, and 12b are STING agonists that are efficacious across species, and all have the skeleton of acridone; 1b, 1c, and 12c just function in the murine STING pathway. Notably, 12b exhibits the best activity among the six agonists, and its inductions of both human and murine STING-dependent signalling are similar to that of 2′3′-cGAMP, which is a well-known STING inducer. While a protein assay indicated that 2 g, 9 g, and 12b could activate the pathway by directly binding human STING, 12b also displayed the strongest binding affinity. Additionally, our studies show that 12b can induce faster, more powerful, and more durable responses of assorted cytokines in a native system than 2′3′-cGAMP. Consequently, our team is the first to successfully modify murine STING agonists to obtain human sensitivity, and these results suggest that 12b is a potent direct-human-STING agonist. Additionally, the acridone analogues demonstrate tremendous potential in the treatment of tumours or viral infections.

Full text of DOI:10.1016/j.bioorg.2019.103556

Journal Pre-proofs
Design, synthesis and biological evaluation of Acridone Analogues as novel
STING receptor agonists
Shi Hou, Xiu-juan Lan, Wei Li, Xin-lin Yan, Jia-jia Chang, Xiao-hong Yang,
Wei Sun, Jun-hai Xiao, Song Li
PII:
S0045-2068(19)31757-2
DOI:
https://doi.org/10.1016/j.bioorg.2019.103556
YBIOO 103556
Reference:
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
18 October 2019
23 December 2019
24 December 2019
Please cite this article as: S. Hou, X-j. Lan, W. Li, X-l. Yan, J-j. Chang, X-h. Yang, W. Sun, J-h. Xiao, S. Li,
Design, synthesis and biological evaluation of Acridone Analogues as novel STING receptor agonists,
Bioorganic Chemistry (2019), doi: https://doi.org/10.1016/j.bioorg.2019.103556
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier Inc.

Design, synthesis and biological evaluation of Acridone Analogues as
novel STING receptor agonists
Shi Hou a,b,c, Xiu-juan Lan a,b,c, Wei Li b,c, Xin-lin Yan b,c, Jia-jia Changb,c,d , Xiao-hong Yang a,*,
Wei Sun a,*, Jun-hai Xiao b,c,*,Song Li b,c
a School of Pharmaceutical Sciences, Jilin University, Changchun 130021, PR China
b
National Engineering Research Center for the Emergency Drug, Beijing Institute of Pharmacology and
Toxicology, Beijing 100850, PR China
c
State Key Laboratory of Toxicology and Medical Countermeasures, Beijing Institute of Pharmacology and
Toxicology, Beijing 100850, PR China
d School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, PR China
* Corresponding author
Xiao-hong Yang: School of Pharmaceutical Sciences, Jilin University, China; E-mail: xiaohongyang88@126.com
Wei Sun: School of Pharmaceutical Sciences, Jilin University, China; E-mail: wsun@jlu.edu.cn
Jun-hai Xiao: National Engineering Research Center for the Emergency Drug, Beijing Institute of Pharmacology
and Toxicology, China; E-mail: xiaojunhai@139.com
ABSTRACT
STING (Stimulator of Interferon Genes) has become a focal point in immunology research and a
target in drug discovery. The discovery of a potent human-STING agonist is expected to
revolutionize current anti-virus or cancer immunotherapy. Inspired by the structure and function of
murine STING-specific agonists (DMXAA and CMA), we rationally designed and synthesized four
series of novel compounds for the enhancement of human sensitivity. In the cell-based assay, we
identified six compounds from all the synthetic small molecules: 2g, 9g, and 12b are STING agonists
that are efficacious across species, and all have the skeleton of acridone; 1b, 1c, and 12c just function
in the murine STING pathway. Notably, 12b exhibits the best activity among the six agonists, and its
inductions of both human and murine STING-dependent signalling are similar to that of

2’3’-cGAMP, which is a well-known STING inducer. While a protein assay indicated that 2g, 9g,
and 12b could activate the pathway by directly binding human STING, 12b also displayed the
strongest binding affinity. Additionally, our studies show that 12b can induce faster, more powerful,
and more durable responses of assorted cytokines in a native system than 2’3’-cGAMP.
Consequently, our team is the first to successfully modify murine STING agonists to obtain human
sensitivity, and these results suggest that 12b is a potent direct-human-STING agonist. Additionally,
the acridone analogues demonstrate tremendous potential in the treatment of tumours or viral
infections.
Keywords: STING; human agonists; direct bonding; induction of cytokines
1. Introduction
STING is both a pattern-recognition receptor (PRR), which directly binds cyclic dinucleotides
(CDNs) via its C-terminal domain [1], and an adaptor molecule, which can be activated by multiple
cytoplasmic receptors of dsDNA [2]. STING plays a vital role in the activation of innate immune
responses. Accumulated evidence has demonstrated that the cyclic GMP-AMP synthase (cGAS) is
the major dsDNA cytosolic sensor allowing STING to function [3, 4]. The essential mechanism for
innate immune sensing of various pathogens and tumours exists in the cGAS–STING pathway in
which cGAS is activated by cytoplasmic DNA to subsequently catalyse the production of cyclic
GMP-AMP (cGAMP). 2’3’-cGAMP is a non-canonical CDN that binds to STING and induces it to
oligomerize [5], leading to a downstream signalling cascade including recruiting the
serine/threonine-protein kinase (TBK1), phosphorylating the interferon regulatory transcription
factor (IRF3) and nuclear factor-κB (NF-κB), and producing cytokines and proteins such as the type
I interferons (IFNs), interleukin 6 (IL-6), and tumour necrosis factor α (TNFα) [6]. IFN-β is the
signature cytokine induced by the host STING pathway, and it primes anticancer activities via
diversified approach including cross-presentation of tumour antigens and mobilization of
tumour-specific CD8 T cells [7]. Therefore, STING is regarded as a promising target in various
fields, including immunization, autoinflammation, and cancer therapeutics [8]. To the best of our
knowledge, current efforts regarding STING agonists are centred on chemically modified CDNs for
improving the polarity and liability of its analogues [9]; some of these analogues strongly and

continually degrade tumours in syngeneic tumour models through intratumoural delivery and are
currently in clinical trials [10]. However, CDNs analogues also exhibit the limitations of low
cell-membrane permeability [11].
Substantial studies to identify more drug-like small molecule STING agonists have been
conducted. According to species selectivity, we classify non-nucleotide STING agonists into three
categories: murine-specific, human-specific, and nonspecific agonists. Since the studies of the
Fitzgerald research team [12] and the Mitchison group [13] were reported, the species-specific of
STING has been attracting great interest. Both of these groups considered that
5′,6′-dimethylxanthenone-4-acetic acid (DMXAA) is a potent agonist of murine STING (mSTING)
rather than human STING (hSTING). These studies clearly explained why DMXAA failed in human
phase III trials against non-small-cell lung cancer (NSCLC). The structure and mechanism of
10-carboxymethyl-9-acridanone (CMA) are similar those of to DMXAA [14]. To the best of our
knowledge, five chemotypes of compounds have been identified as the human STING-specific
agonist [15-19]. Except for a dietary xanthone α-mangostin that directly binds and activates human
STING, other small molecules share the following properties: indirect STING agonists cannot bind
to STING, identification of chemical compounds by cell-based high-throughput screening (HTS),
comprehensive and diverse bioactivities were lower than that of 2’3’-cGAMP in vitro, and the
existence of unknown difficulties in animal experiments. In 2018, 44 scientists jointly published a
new study in Nature. These scientists used STING competition binding assays to screen small
molecules and found an aminobenzimidazole (ABZI) compound and its dimer (di-ABZIs)
derivatives as potent and nonspecific STING agonists [20]. Although di-ABZIs have shown
pre-clinical promise as STING agonists, the relatively large molecular weight and polarity, which are
similar to those of 2’3’-cGAMP, probably limit the application for intratumoural injection. Analysis
of various STING agonists above, the further effort should be focused on the identification of small
molecules STING agonists that are efficacious across species. Furthermore, indistinct mechanisms of
the STING pathway may be elucidated by them as tool molecules, and their application in animals
enables to facilitate the assessment of safety.
2. Results and Discussion

2.1 Design and Synthesis of xanthenone-4-acetic acid（XAA）and acridone Analogs
Undeniably, murine STING activators DMXAA and CMA have some inherent advantages: 1)
They are small molecular compounds, which can mitigate the defects of CDNs [21]; 2) Both of them
can bind to murine STING protein, and have confirmed crystal structures [14, 22]; 3) The seminal
contributions of Patel and co-workers, which includes subtle mutations in protein sequences of
STING now able to endow human STING with DMXAA sensitivity [23], inspire us to begin our
study. Hence, our innovations of molecule design originate from comprehensively analysing the
similarities of DMXAA and CMA, especially the relationship between molecules and proteins
(Figure 1A-C).
Figure 1. Intermolecular contacts in the complex of small molecules bound to STING with the representation of
hydrogen bonding as the dotted line. (A) CMA bound to murine STING [14]; (B) DMXAA bound to murine
S162A/Q266I
STING [22]; (C) DMXAA bound to human STING
[23]; (D) The corresponding location between

DMXAA and mutant residues.
The following information has been revealed: maintaining the structure of three rings
guarantees the interaction between two small molecules; the 5,6-dimethyl substitutes of DMXAA are
highly significant since it can produce a strong hydrophobic effect with the surrounding amino acids
(Figure 1D); the mutant substitutions (S162A and Q266I) of hSTING at the region of
binding-pocket correspond to the positions of DMXAA at C1/C2 and C7 (Figure 1D). Previous
work synthesized C7-position DMXAA derivatives that contain a hydrogen bonding donor and
acceptor or a halide, but none of the compounds can bind to hSTING [24]. Thus, we designed and
synthesized four types of small molecules of a relevance skeleton whose changes concentrated on
assorted groups at C1/C2 and C7 positions of compounds. The structures and synthetic methods of
design molecules are summarized in Figure 2 and Scheme 1–4.
Figure 2. The structures of design molecules
The DMXAA synthetic protocol could be applied to our synthesis of series Ⅰ (Scheme 1) [25].

Cross-coupling between the sodium phenolate 5a, 5b and potassium salt of Iodide 6a, 6b followed
by a Friedel–Crafts acylation-type reaction produced DMXAA and other XAAs (1a–1d). The
synthetic route for designed compounds 2a–2o (series Ⅱ) is outlined in Scheme 2 [26]. Thus,
Ullmann condensation between phenylacetic acid derivative 7a–7e and 5-substituted-2-amino-3,4-
dimethyl-benzoic acid 8a–8c in the presence of Cu powder, Cu2O and K2CO3 in DMF, and synthetic
compounds of benzene acetic acid were treated with Eaton’s reagent to afford the cyclized
intermediate products. End-products were obtained by opening five-membered rings under alkaline
condition. In Scheme 3, we show the synthesis of a series of multi-site-substituted acridones (series
Ⅲ) [27]. First, acridone analogous 12a–12f could be accessed using Scheme 2 synthetic procedures.
Subsequent ester compounds 13a, 13b were synthesized by coupling ethyl bromoacetate [28], and
we prepared the final compounds 3a, 3b by hydrolysis and acidification. Ullmann’s reaction and
Friedel-Crafts acylation were also carried out in Scheme 4 to give another kind of acridone
derivative (series Ⅳ): 2-methoxy-4-carboxylic acid acridone 4. The structures of the compounds
were listed in Table 1 and confirmed on the basis of MS and NMR (1H and 13C) spectral data.
However, the number of molecules was significantly limited by the precise modification sites and the
commercial availability of the origin of benzene derivatives.
O
KOOC
R1
R1
a,b
COONa
ONa
5a,5b
I
R3
R3
O
R2
R2
COOH
6a,6b
1a-1d
Scheme 1. The synthetic routes of 1a-1d (series Ⅰ). (a) DMSO, CuCl, TDA-1, 95°C; (b) H2SO4, H2O (9:1), 80°C.
O
R1
O
R1
R3
R2
HOOC
H2N
R3
R3
R2
R2
R1
a,b
c
COOH
N
O
N
H
Br
COOH
2a-2o
9a-9o
7a-7e
8a-8c
Scheme 2. The synthetic routes of 2a–2o (series Ⅱ). (a) DMF, Cu, Cu 2 O, K2CO3, 100-120 °C; (b) Eaton’s reagent, 80-100 °C; (c)
EtOH, H2O, NaOH, reflux.
O
N
O
N
O
R4
R3
R1
R4
R3
R1
R4
R3
a,b
R4
R3
R1
c
d
R1
COOH
Br
H2N
N
H
R2
O
R2
OH
R2
R2
O
O
3a,3b
10a,10b
11a-11c
12a-12f
13a,13b

Scheme 3. The synthetic routes of 3a, 3b (series Ⅲ). (a) DMF, Cu, Cu 2 O, K2CO3, 100-120 °C; (b) Eaton’s reagent, 80-100 °C; (c)
BrCH2 COOC2H5, NaH, r.t.; (d) NaOH, 2 M HCl.
O
HOOC
a,b
O
COOH
Br
O
H2N
N
H
COOH
10b
8a
4
Scheme 4. The synthetic routes of 4 (series Ⅳ). (a) DMF, Cu, Cu 2 O, K2CO3, 100-120 °C; (b) Eaton’s reagent, 80-100 °C.
2.2 In Vitro Screening of STING Agonists and the Structure−Activity Relationship (SAR) analysis.
To thoroughly evaluate the four kinds of analogues, we elaborately selected three cell lines in
which the STING gene has been transfected or knocked out: 293T-hSTING-R232 cells,
293T-mSTING cells, and THP1-KO-STING cells. STING is either deactivated, undetectable, or not
expressed in certain cell lines, such as human embryonic kidney cells (HEK293) [29]. However,
293T-STING cells are a family of reporter cells designed to study variants of STING. Recent work
revealed that the sequence differences among these variants can markedly affect STING function
and, consequently, impact human health. Our choice of 232R-RGR (71R-230G-293R) is the most
prevalent variant known as the wild-type STING gene and accounts for approximately 70% of the
human population [30]. Therefore, human or murine specific STING agonist stimulation can be
respectively assessed in 293T-hSTING-R232 cells or 293T-mSTING cells by monitoring
IFN-stimulated response elements and (ISRE)-induced secreted embryonic alkaline phosphatase
(SEAP) production. Then, we employed THP1-KO-STING cells to confirm whether the compound
exhibits the function of STING-dependent cytokine induction.
The screening process was divided into two steps. First, we used three STING cell lines to screen
all targeted compounds and their synthetic precursors at two concentrations (100 μM, 20 μM) to
quickly identify small molecules that induced I IFN via the STING pathway. The structures of the
screening compounds were displayed in Figure 3, and the first screened results were listed in Table
1.

O
O
O
O
R1
R1
R1
O
R3
R2
R3
O
R3
N
N
H
4
N
H
2a-2o
R2
R2
COOH
COOH
O
COOH
1a-1d
COOH
3a,3b
O
O
O
O
R4
R3
R1
N
O
N
N
H
O
R2
12a-12f
9g
13b
O
Figure 3. The structure of screening small molecules
Table 1
Structures and Biological grades of the XAA and acridone Analogues of the Preliminary Screening
Compound
R1
H
Br
H
Br
H
F
R2
CH3
CH3
H
R3
CH3
CH3
OCH3
OCH3
H
R4
-
Human-STINGa Murine-STINGb
1a(DMXAA)
-
-
-
-
-
-
-
-
-
-
+
-
-
-
-
-
-
-
++
+
+
-
1b
1c
1d
2a
2b
2c
2d
2e
2f
-
-
H
-
H
-
-
H
H
-
-
Cl
H
H
H
H
F
H
H
-
-
F
H
-
-
Cl
H
H
-
-
Br
-
-
2g
2h
2i
H
OCH3
Br
-
+
-
H
-
Cl
H
H
F
H
Br
-
-
2j
F
Br
-
-
2k
2l
Cl
H
Br
-
-
OCH3
OCH3
OCH3
-
-
2m
2n
Cl
H
H
-
-
F
-
-

2o
3a(CMA)
3b
H
H
Cl
H
OCH3
H
-
-
-
-
+
-
-
OCH3
-
CH3
-
CH3
-
-
-
4
-
-
-
9g
H
H
OCH3
H
-
H
+
-
+
-
12a
12b
12c
12d
12e
12f
H
H
OCH3
OCH3
H
CH3
H
CH3
H
H
++
-
++
+
-
H
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
-
H
OCH3
OCH3
H
-
-
OCH3
OCH3
-
-
13b
-
-
Measuring the OD (optical density) value of a small molecule at two concentrations (100 μM, 20 μM) and
comparing with 2’3’-cGAMP (30 μg/ml), we classified the activity of small molecules into three grades: “++”,
excellent (the OD value of the compound at 20 μM is approximately the 2’3’-cGAMP); “+”, good (the OD value of
the compound at 100 μM is approximately the 2’3’-cGAMP); “-”, inactive (the OD value of the compound at 100
μM is much lower than the 2’3’-cGAMP). aEach result does three independent experiments in 293T-hSTING-R232
cells. bEach result does three independent experiments in 293T-mSTING cells.
Secondly, we used differentiated concentration gradients to test and verify the intensity of
screened small molecules and compared their activities to that of the reference STING agonists for
murine (DMXAA) and human (2’3’-cGAMP). The function of agonists in STING-dependent
cytokine induction has been confirmed by using THP1-KO-STING cells. Together, six potent
compounds were obtained from the various synthetic precursors and their derivatives by two-step
selection. Three were non-species-specific agonists (2g, 9g, and 12b), and none of these compounds
exhibited any IRF or NF-κB activities in the knockout hSTING cells, while the other compounds
(1b, 1c, and 12c) just activated the murine STING pathway (Figure 4 and Table 2).

O
O
O
Br
O
O
O
O
N
H
12c
1b
1c
HOOC
O
HOOC
O
O
O
N
O
O
N
H
N
H
COOH
O
9g
2g
12b
Figure 4. The structures of screened agonists.
Table 2
Biological Activities of the screened agonists of the Second Screening and of DMXAA and 2’3’-cGAMP a, b
Human-STING
Murine-STING
EC50 (μM)a
Compound
EC50 (μM)a
Emax (%)b
Emax (%)b
63.15±6.5
69.64±3.4
78.32±5.6
83.39±4.1
125.91±20.4
60.45±5.3
118.27±15.8
82.39±10.4
100
1b
nt
nt
nt
152.91±6.6
115.59±6.6
66.45±10.5
72.33±11.1
10.23±0.6
134.62±5.9
8.87±1.1
1c
2g
nt
154.21±9.6
134.59±12.1
7.45±2.7
nt
63.8±8.8
70.3±8.5
109±19.7
nt
9g
12b
12c
DMXAA(1a)
CMA(3a)
2’3’-cGAMP
nt
nt
nt
nt
96.21±6.6
7.23±2.6
100
7.01±2.1
a
b
Each value is the mean ± SD from three independent experiments at least. Emax values are reported as a mean
percentage ± SD relative to the corresponding 2’3’-cGAMP response from three independent experiments at least.
In our study, the refinements of DMXAA also proved to be unsuccessful. Synthesized small
molecules not only fail to trigger human STING pathway induction but also negatively respond to
the mouse pathway with expanding or changing substituents in DMXAA. Fortunately, some of the
acridone analogues we discovered were powerful and helpful, especially compound 12b, which
displayed approximately similar activities to 2’3’-cGAMP at activating both pathways (human:
EC50=7.45 μM, murine: EC50=10.23 μM).
In acridone analogues, we mainly investigated the effects of the different substitutions at various

positions and explored and obtained some preliminary SARs as much as possible. All the screened
human-STING agonists, which are typical acridone structures, consist of two benzene rings fused
together with a keto-group and a nitrogen atom at the 9th and 10th position, exhibit the common
substitutes of 5,6-dimethyl and verify the rationality of our plan. For the series Ⅱ compounds, we
found that the modifications of the C1/C2 position were ineffective. But, 2g and 9g (the series Ⅱ
compounds of C7-modification) shown better activities than CMA in murine-STING derived cells,
and had STING stimulating activities in human-STING derived cells. It indicated that the methoxyl
at the C7-position was crucial for inducing the STING pathway. By analysis of series Ⅲ and Ⅳ
compounds, we discovered that their intermediates without carboxylic acid structure are active.
Specifically, the synthetic precursors 12b differs from compounds 3b and 4 in the absence of
corresponding carboxylic acids at N or C-4 position, and the addition of the substituent enormously
shrank the activity of the molecules. Meanwhile, we compared 12b, 12c, and 12d noticed that the
importance of the C-2-methoxyl is the key group in small molecules to the activation of the STING
pathway. The structure of 5,6-dimethyl is to endow acridone compounds with the sensitivity of
human STING (Figure 5).
Figure 5. Summarized SARs of the acridone analogues.

Interestingly, compound 12e, which maintains the structure of 5,6-dimethyl-acridone and only
substitutes methoxyl at the C-7 position, is inactive and indicates that the C-7-methoxyl may work in
coordination with groups of other positions (maybe at N and C-4). The active response of acridones
increases when C-2 or C-7 is substituted by methoxyl; however, the disappearance in the induction
of type I IFNs after two sites are replaced together (reference compound 12f). Accordingly, for the
future design of acridone agonists, we need to maintain the structure of 12b and decorate the
positions of N, C-4 and C-7 with suitable groups to strengthen the activity in vitro.
2.3 The Acridone Agonists Directly Bind and Stabilize the hSTING Protein
Next, we explored the reason why acridones agonists vigorously stimulated the STING pathway
and further evaluated the biological activity of them. To verify whether acridones agonists directly
activate the STING-TBK1-IRF3 pathway through STING, we first used human STING binding kits
provided by Cisbio. Using a homogeneous time-resolved fluorescence (HTRF) competition format,
WT
which uses a STING ligand-d2, a 6His tagged human STING
protein, and an anti-6His
Cryptate-labelled antibody, STING binders compete with the STING ligand-d2, thus preventing
fluorescence resonance energy transfer (FRET) from occurring. The competitive binding data
identified 12b (IC50=1.92 μM) as exhibiting the strongest ligand-human STING interaction among
all our synthetic agonists. The IC50 value of 2’3’-cGAMP was 4.6 nM in agreement with the
reference values provided by kits Instructions(2’3’-cGAMP Standard IC50=5 nM), and DMXAA
(purple rhombuses, IC50 > 500 μM), as a negative control drug, does not compete with the binding
of STING ligand-d2 to the human STING protein as expected (Figure 6A).
To further cross-validate the direct binding of 12b with hSTING CTD, we conducted a universal
and authoritative method called isothermal titration calorimetry (ITC) to measure the Kd value of
12b. The results of ITC were consistent with that of the binding kit, and the Kd values of
2’3’-cGAMP and 12b were 0.02 μM and 26.4 μM, respectively (Figure 6B-D).

Figure 6. The acridone agonists bind and stabilize the hSTING CTD. (A) STING competition binding assay.
Dose-response curves of 2’3’-cGAMP (n = 3), 12b (n = 4), 2g (n = 3), 9g(n=3), DMXAA(n=3), Error bars
represent SD of independent experiments. Relative potency of compounds was calculated using GraphPad Prism
software, Mean response ± SD. The interaction of hSTINGWT CTD with buffer (B), compound 12b (C) and
2’3’-cGAMP (D) was quantified using isothermal titration calorimetry (ITC).
The great differences in binding affinity between 12b and 2'3'-cGAMP could be explained by
their crystal structures of binding STING [31]: 2’3’-cGAMP bound to the symmetrical dimer of
human STING by a single molecule (Figure 7A), but there are two molecules of DMXAA in their
complex with the dimer of murine STING (Figure 7B). The A rings of the two DMXAA molecules
overlapped each other (the distance was 3.9 Å) to form a π – π stacking interaction that kept the

structure of STING dimer (Figure 7C). For the weak interaction of the π – π stacking interaction, the
binding force of DMXAA was much lower than that of 2’3’-cGAMP. Hornung research group also
reported that the acridone ring moieties of both CMAs partially stack to each other in a parallel (~ 4
Å distance) [14]. Therefore, we speculated that the mechanism of compound 12b is the same as that
of DMXAA and CMA.
Figure 7. The binding mechanisms of STING agonists. (A) The 2.25 Å crystal structure of 2’3’-cGAMP bound to
H232
hSTING
(PDB code: 4LOH); (B) The 2.43 Å crystal structure of two molecules of DMXAA bound to
mSTING R231 (PDB code: 4LOL); (C) The distance between the A rings of the two DMXAA molecules (PDB code:
4LOL).
Similarly, the Kd values of natural STING-ligands 2'3'-cGAMP and 3'2'-cGAMP are nearly 1000
times different, but their activation levels of IFN-β are equipotent in the micro molar range[3].
Furthermore, the binding affinity is not necessarily related to the activation of small molecules in the
STING pathway, and small molecules with high affinity may be the STING inhibitors. Significantly,
different types of agonists have different binding mechanisms, the mechanism of compound 12b
needs to be further determined. Furthermore, we have successfully demonstrated that acridone

agonists can directly bind to human STING. The binding affinity of acridone derivatives are exactly
proportional to cellular effects in vitro, and this reveals that the point of design ideas about
modification of this kind of derivatives can be used to enhance the affinity between small molecules
and STING C-terminal domain (CTD).
2.4 Identification of an Acridone Compound That Potently Activates the cGAS−Human STING
Pathway
To confirm the functional consequence of binding, we employed real-time PCR (qPCR) to
examine the effects of the best compound 12b on proinflammatory-cytokine gene expression in a
native system (THP-1 cells), using well-known STING binders 2’3’-cGAMP as a control. Firstly, we
recognized the fitting concentration of 12b (40 μM) through preliminary experiments. Then, the
experiments of time gradient (co-cultivation time) were used to further weigh the competence of two
agonists in secreting cytokines (Figure 8). Interestingly, the advantage of the 2’3’-cGAMP lied only
in the expression of IP-10, which is an interferon-stimulating gene, while the maximum expression
of other immune factors induced by 12b were greater than that induced by 2’3’-cGAMP, especially
TNF-α. Considering the impact of time, the mRNA level of most cytokines was enhanced with time
by 12b and 2’3’-cGAMP, except for the induction of IFN-β by 12b, which showed a slight
downward trend.

Figure 8. 2’3’-cGAMP and 12b induces diverse cytokines production in a native system. THP1 cells were
stimulated with 21 μM 2’3’-cGAMP or 40 μM 12b. mRNA level of IFNβ (A), TNFα (B), IL-6 (C) and IP-10(D)
were measured at 8h, 20h and 30h by real-time PCR and normalized by GAPDH expression. Error bars represent
SEM of independent experiments (n=3). Significance (* Pꢀ<ꢀ0.05, **ꢀPꢀ<ꢀ0.01, ***ꢀPꢀ<ꢀ0.001) was determined by
t-test relative to DMSO at a particular time.
By comprehensive analysis, we observed that 12b and 2’3’-cGAMP generally exhibited potent
action in excreting cytokines. Although the former gave a more fast, powerful, and durable
induction, the results can be explained by the different mechanisms between 12b and 2’3’-cGAMP.
STING agonists can activate two pathways: the STING-IRF and the STING-NF-κB. STING-IRF3
pathway mainly induces the production of type Ⅰ interferons, and the pro-inflammatory cytokines
are secreted by the STING-NF-κB pathway. The production of TNF-α is the marker in
NF-κB-pathway [32]. We boldly presumed that 12b can induce both STING pathways, and the
activation of NF-κB-pathway is more active than 2’3’-cGAMP.
3. Conclusions
The past few years have witnessed the rapid advance of the novel STING agonists as
anti-tumour, anti-virus, and immunostimulatory agents. However, the best agonists of CDNs

analogues were also confined to poor permeability of cell membranes in intra-tumoural drug delivery
of clinical trials [11]. The present research aimed to discover novel STING agonists that may result
in more drug-like properties and lead to a breakthrough at species-specificity. A promising thought is
to modify the mouse-specific STING agonist DMXAA, but no derivatives have been identified in
binding human STING to date [24]. To overcome these challenges, we summarized the general
character of mouse-specific STING agonists (DMXAA and CMA), rationally designed and
synthesized various novel XAA and acridone derivatives, and investigated the biological activity of
them. Notably, compound 12b is a non-species-specific STING agonist with an acridone skeleton,
exhibiting the most potent activity with the induction of type I interferons among all synthetic
compounds. The levels of the activation of STING signalling in vitro are similar to those of the
2’3’-cGAMP. Additionally, some information provided by comprehensive and detailed SARs
implied that removing any substituents of 12b could weaken its activity. Therefore, we need to
preserve the structure and change other sites for improvement of the effects. We further
demonstrated that 12b directly binds human STING. The binding capability between the acridone
analogues and STING is tightly associated with its activity in vitro. In our study of the mechanism,
showing its ability to lead to vigorous activation in both STING pathways of IRF3 and NF-κB, 12b
exhibits the tremendous potential to be a better direct-human-STING agonist than 2’3’-cGAMP.
Accordingly, improving the activity or drug properties of acridone analogues still results in many
possibilities, such as adding suitable groups at positions N, C-4 and C-7 or synthesising the
double-structured molecules by imitating di-ABZIs in order to increase the binding affinity between
the molecule and STING.
4. Experiment Section
4.1. Chemistry
1
All reagents and solvents were used as received from commercial sources. H-NMR and
13C-NMR spectra were recorded at 400 MHz and 100 MHz on a JNM-ECA-400 instrument (JEOL
Ltd., Tokyo, Japan) in DMSO-D6. Chemical shifts are expressed in δ (ppm), with tetramethyl silane
(TMS) functioning as the internal reference. Coupling constants (J) were expressed in Hz.
High-resolution mass spectra were obtained using a TOF G6230A LC/MS (Agilent Technologies,

New York, NY, USA) with an ESI source. Reagents and solvents were commercially available
1
13
without further purification. The H-NMR, C-NMR and HRMS spectra of the compounds in this
article can be found in Supporting Information.
5′,6′-dimethylxanthenone-4-acetic acid (DMXAA-1a).
2-hydroxyphenylacetic acid 5a was
prepared to disodium salt and 2-iodo-3,4-dimethylbenzoic acid 6a was converted into its potassium
salt. To a solution of 5a (0.62 g, 3.2 mmol), 6a (1 g, 3.2 mmol), TDA-1 (0.2 g, 0.64 mmol) and CuCl
(0.06g, 0.64 mmol) in dry DMSO (5ml) was added. Then the mixture was stirred at 85°C under
nitrogen for 8 h. Upon cooling to room temperature and removal of the solvents under a vacuum.
The residue was dissolved with 1N NaOH (20ml) followed by filtration. The solvent was acidified to
PH=7 by slowly dropping the conc.HCl before filtered to give crude 2-(2-(carboxymethyl)
phenoxy)-3,4-dimethylbenzoic acid (0.73g, 76.0% yield). After the first step, the intermediate
product was cyclodehydrated in conc.H2SO4 (4 ml) and water (1 mL) at 100°C for 2h. After pouring
the reaction mixture into ice cold water (25ml), the crude 1a was collected by filtration and was
crystallised from EtOAc/MeOH to give pale white solid DMXAA (0.58 g, 64.4% yield over two
steps). 1H NMR (400 MHz, DMSO-D6) δ (ppm): 8.06 (dd, J = 8.0, 1.6 Hz, 1H), 7.87 (d, J = 8.1 Hz,
1H), 7.77 (dd, J = 7.2, 1.5 Hz, 1H), 7.39 (t, J = 7.6 Hz, 1H), 7.22 (d, J = 8.1 Hz, 1H), 3.94 (s, 2H),
2.38 (s, 3H), 2.34 (s, 3H). 13C NMR (101 MHz, DMSO-D6) δ (ppm): 176.62, 172.41, 154.26,
153.74, 145.20, 137.10, 126.48, 125.82, 125.64, 125.13, 124.15, 123.11, 121.01, 119.26, 35.98,
20.71, 11.59. HRMS (ESI) m/z [M+ H] + calculated for C17H14O4: 283.09 found: 283.09.
(2-bromo-5,6-dimethyl-9-oxo-9H-xanthen-4-yl)-acetic acid (1b).
The preparation of title
compounds was similarly to 1a. 2-hydroxyphenylacetic acid 5a was replaced with
1
2-(5-bromo-2-hydroxyphenyl) acetic acid 5b. (pale gray solid, yield: 51.3%). H NMR (400 MHz,
DMSO-D6) δ (ppm): 12.72 (s, 1H), 8.14 (d, J = 2.5 Hz, 1H), 8.03 (d, J = 2.5 Hz, 1H), 7.92 (d, J = 8.3
13
Hz, 1H), 7.32 (d, J = 8.2 Hz, 1H), 4.00 (s, 2H), 2.44 (s, 3H), 2.39 (s, 3H). C NMR (101 MHz,
DMSO-D6) δ (ppm): 180.62, 176.41, 158.26, 157.74, 149.20, 141.10, 137.33, 130.48, 129.82,
129.13, 128.15, 127.11, 125.01, 120.19, 39.98, 24.71, 15.59. HRMS (ESI) m/z [M+ H] + calculated
for C17H13BrO4: 362.99 found: 363.00.

(6-methoxy-9-oxo-9H-xanthen-4-yl)-acetic acid (1c).
The title compound was obtained similarly
to 1a. 2-iodo-3,4-dimethylbenzoic acid 6a was replaced with 2-iodo-4-methoxybenzoic acid 6b.
1
(yellow solid, yield: 45.8%). H NMR (400 MHz, DMSO-D6) δ (ppm): 12.59 (s, 1H), 8.12 – 8.07
(m, 2H), 7.78 (dd, J = 7.3, 1.7 Hz, 1H), 7.42 (t, J = 7.9, 7.3 Hz, 1H), 7.11 – 7.04 (m, 2H), 3.98 (s,
2H), 3.94 (s, 3H). 13C NMR (101 MHz, DMSO-D6) δ (ppm): 175.49, 172.49, 165.48, 157.78,
154.39, 136.95, 128.07, 125.29, 125.19, 124.28, 121.62, 115.17, 114.49, 100.84, 56.70, 35.20.
HRMS (ESI) m/z [M+ H] + calculated for C16H12O5: 285.07 found: 285.08.
(2-bromo-6-methoxy-9-oxo-9H-xanthen-4-yl)-acetic acid (1d).
The preparation of title compound
was similarly to 1a. 2-hydroxyphenylacetic acid 5a and 2-iodo-3,4-dimethylbenzoic acid 6a were
replaced with 2-(5-bromo-2-hydroxyphenyl) acetic acid 5b and 2-iodo-4-methoxybenzoic acid 6b
1
respectively. (brown solid, yield: 40.3%). H NMR (400 MHz, DMSO-D6) δ (ppm): 12.79 (s, 1H),
8.10 (d, J = 2.1 Hz, 1H), 8.05 (d, J = 8.8 Hz, 1H), 7.98 (s, 1H), 7.09 – 7.01 (m, 2H), 3.97 (s, 2H),
13
3.93 (s, 3H). C NMR (101 MHz, DMSO-D6) δ (ppm): 174.25, 172.00, 165.73, 157.67, 153.43,
139.06, 128.36, 128.20, 127.04, 122.98, 116.34, 114.84, 114.63,100.90, 56.77, 34.73. HRMS (ESI)
m/z [M+ H] + calculated for C16H11BrO5: 362.98 found: 362.99.
General Procedure for the Preparation of (5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic
acid (2a).
To a magnetically stirred solution of 2-(2-bromophenyl)acetic acid 7a (0.72g,
3.3mmol) and 2-amino-3,4-dimethylbenzoic acid 8a (0.5g, 3.0mmol) in DMF was added powdered
Cu (0.02g), Cu2O (0.02g) and K2CO3 (0.33g, 2.4mmol) successively. The reaction mixture was
heated at 95℃ for 12h. Upon removal of the solvents, the residue was triturated with 1N NaOH
solution. The whole mixture was filtered off and the alkaline solution was acidified with conc.HCl,
and then collected the precipitate. The dried residue was subjected to column purification (PE/EA
2:1) to furnish the desired compound. A solution of 2-((2-(carboxymethyl) phenyl)
amino)-3,4-dimethylbenzoic acid in Eaton’s reagent (P2O5 –CH3SO3H) (5 mL) at room temperature
was heated to 90 °C for 1 h. After cooling to room temperature, the mixture was added slowly to the
saturated aqueous NaHCO3 solution. The crude 9,10-dimethyl-6H-pyrrolo[3,2,1-de]
acridine-1,6(2H)-dione was accumulated by filtration followed by purification with silica gel column

chromatography.
Finally,
NaOH
(0.1g)
was
added
to
a
solution
of
9,10-dimethyl-6H-pyrrolo[3,2,1-de] acridine-1,6(2H)-dione in EtOH (20 mL) and water (2ml). The
resulting mixture was refluxed overnight. Upon removal of the solvents, the residue was dissolved
with water. Then, the solution was neutralized by conc.HCl. After filtering the mixture, the crude
was obtained and purified by silica gel column chromatography to give 0.36g 2a as yellow solid in a
yield of 43.8%. 1H NMR (400 MHz, DMSO-D6) δ (ppm): 13.14 (s, 1H), 9.50 (s, 1H), 8.19 (dd, J =
8.1, 1.3 Hz, 1H), 8.02 (d, J = 8.2 Hz, 1H), 7.69 (dd, J = 7.2, 1.5 Hz, 1H), 7.25 (m, J = 8.0, 7.2 Hz,
1H), 7.16 (d, J = 8.3 Hz, 1H), 4.12 (s, 2H), 2.50 (s, 3H), 2.45 (s, 3H). 13C NMR (101 MHz,
DMSO-D6) δ (ppm): 177.54 (s), 173.73 (s), 142.42 (s), 139.74 (s), 139.48 (s), 135.94 (s), 125.89 (s),
124.65 (s), 123.64 (s), 123.17 (s), 121.67 (s), 121.26 (s), 119.43 (s), 38.77 (s), 21.21 (s), 12.70 (s).
HRMS (ESI) m/z [M+ H] + calculated for C17H15NO3: 282.11 found: 282.11.
(1-fluoro-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic acid (2b).
The preparation of
title compound was similarly to 2a. 2-(2-bromophenyl) acetic acid 7a was replaced with
2-(2-bromo-4-fluorophenyl) acetic acid 7b. The pale gray solid was got, yield: 32.8%. 1H NMR (400
MHz, DMSO-D6) δ (ppm): 13.17 (s, 1H), 9.43 (s, 1H), 7.95 (d, J = 8.2 Hz, 1H), 7.65 (m, J = 8.2, 5.4
Hz, 1H), 7.15 (d, J = 8.2 Hz, 1H), 6.95 (m, J = 11.7, 8.2 Hz, 1H), 4.07 (s, 2H), 2.47 (s, 3H), 2.44 (s,
3H). 13C NMR (101 MHz, DMSO-D6) δ (ppm): 198.95, 196.39, 185.29, 182.70, 165.35, 164.12,
161.50, 158.98, 147.73, 146.17, 145.85, 143.32, 141.64, 130.24, 61.10, 43.87, 35.35. HRMS (ESI)
m/z [M+ H] + calculated for C17H14FNO3: 300.10 found: 300.10.
(1-chloro-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic acid (2c).
The title compound
was obtained similarly to 2a. 2-(2-bromophenyl) acetic acid 7a was replaced with
1
2-(2-bromo-4-chlorophenyl) acetic acid 7c. (pale red solid, yield: 28.5%). H NMR (400 MHz,
DMSO-D6) δ (ppm): 13.28 (s, 1H), 9.53 (s, 1H), 7.95 (d, J = 8.2 Hz, 1H), 7.59 (d, J = 7.9 Hz, 1H),
7.23 (d, J = 7.9 Hz, 1H), 7.15 (d, J = 8.3 Hz, 1H), 4.09 (s, 2H), 2.46 (s, 3H), 2.43 (s, 3H). 13C NMR
(101 MHz, DMSO-D6) δ (ppm): 176.71, 173.48, 142.51, 142.02, 138.42, 135.38, 132.64, 125.03,
124.25, 123.73, 123.00, 122.71, 120.71, 117.10, 38.80, 21.10, 12.55. HRMS (ESI) m/z [M+ H] +
calculated for C17H14ClNO3: 316.07 found: 316.07.

(2-fluoro-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic acid (2d).
The preparation of
title compound was similarly to 2a. 2-(2-bromophenyl) acetic acid 7a was replaced with
1
2-(2-bromo-5-fluorophenyl) acetic acid 7d. The brown solid was obtained, yield: 37.6%. H NMR
(400 MHz, DMSO-D6) δ (ppm): 13.35 (s, 1H), 9.88 (s, 1H), 8.00 (d, J = 8.3 Hz, 1H), 7.82 (dd, J =
8.9, 3.1 Hz, 1H), 7.68 (dd, J = 8.9, 3.0 Hz, 1H), 7.16 (d, J = 8.4 Hz, 1H), 4.16 (s, 2H), 2.48 (s, 3H),
13
2.44 (s, 3H). C NMR (101 MHz, DMSO-D6) δ (ppm): 199.60, 196.05, 180.90, 178.51, 165.30,
162.29, 159.54, 149.63, 147.61, 146.80, 146.22, 144.47, 141.40, 131.91, 61.72, 43.93, 35.50. HRMS
(ESI) m/z [M+ H] + calculated for C17H14FNO3: 300.10 found: 300.10.
(2-chloro-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic acid (2e).
The title compound
was obtained similarly to 2a. 2-(2-bromophenyl) acetic acid 7a was replaced with
1
2-(2-bromo-5-chlorophenyl) acetic acid 7e. (pale brown solid, yield: 39%). H NMR (400 MHz,
DMSO-D6) δ (ppm): 13.78 (s, 1H), 11.00 (s, 1H), 7.99 (d, J = 2.3 Hz, 1H), 7.97 (s, 1H), 7.55 (d, J =
2.5 Hz, 1H), 7.10 (d, J = 8.3 Hz, 1H), 3.74 (s, 2H), 2.50 (s, 3H), 2.41 (s, 3H). 13C NMR (101 MHz,
DMSO-D6) δ (ppm): 176.68, 173.66, 142.04, 140.82, 140.22, 133.79, 130.29, 125.36, 124.55,
124.48, 123.42, 122.81, 121.92, 119.49, 45.15, 21.23, 13.49. HRMS (ESI) m/z [M+ H] + calculated
for C17H14ClNO3: 316.07 found: 316.07.
(7-bromo-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic acid anion (2f).
The title
compound was obtained similarly to 2a. 2-amino-3,4-dimethylbenzoic acid 8a was replaced with
2-amino-5-bromo-3,4-dimethylbenzoic acid 8b. The starting material 8b was synthesized according
to a method previously reported in the literature.[25] The dark brown solid was provided in 28.8%
yield. 1H NMR (400 MHz, DMSO-D6) δ (ppm): 9.69 (s, 1H), 8.25 (s, 1H), 8.17 (d, J = 8.1 Hz, 1H),
7.71 (d, J = 7.1 Hz, 1H), 7.27 (t, J = 7.6 Hz, 1H), 4.12 (s, 2H), 2.58 (s, 3H), 2.54 (s, 3H). 13C NMR
(101 MHz, DMSO-D6) δ (ppm): 176.46, 173.80, 140.98, 139.76, 138.55, 136.29, 126.81, 126.18,
125.80, 123.50, 122.08, 121.15, 120.14, 118.71, 38.96, 21.10, 14.32. HRMS (ESI) m/z [M+ H] +
calculated for C17H14BrNO3: 360.02 found: 360.02.

(7-methoxy-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic acid (2g).
The preparation of
title compound was similarly to 2a. 2-amino-3,4-dimethylbenzoic acid 8a was replaced with
2-amino-5-methoxy-3,4-dimethylbenzoic acid 8c. The starting material 8c was synthesized
1
according to a method previously reported in the literature.[25] (gray solid, yield: 25%). H NMR
(400 MHz, DMSO-D6) δ (ppm): 9.49 (s, 1H), 8.19 (dd, J = 8.1, 1.3 Hz, 1H), 7.67 (dd, J = 7.2, 1.3
Hz, 1H), 7.53 (s, 1H), 7.23 (dd, J = 8.0, 7.2 Hz, 1H), 4.11 (s, 2H), 3.89 (s, 3H), 2.53 (s, 3H), 2.32 (s,
3H). 13C NMR (101 MHz, DMSO-D6) δ (ppm): 176.85, 173.75, 153.25, 139.29, 135.45, 134.39,
133.16, 125.81, 125.23, 123.10, 121.36, 120.67, 119.38, 101.69, 55.97, 38.78, 13.40, 13.26. HRMS
(ESI) m/z [M+ H] + calculated for C18H17NO4: 312.12 found: 312.14.
(7-bromo-1-fluoro-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic
acid
(2h).The
title
compound was obtained similarly to 2a. 2-(2-bromophenyl) acetic acid 7a and
2-amino-3,4-dimethylbenzoic acid 8a were replaced with 2-(2-bromo-4-fluorophenyl) acetic acid 7b and
1
2-amino-5-bromo-3,4-dimethylbenzoic acid 8b respectively. (dark yellow solid, yield: 17.3%). H
NMR (400 MHz, DMSO-D6) δ (ppm): 13.23 (s, 1H), 9.55 (s, 1H), 8.17 (s, 1H), 7.67 (dd, J = 8.2, 5.4
13
Hz, 1H), 6.97 (dd, J = 11.5, 8.2 Hz, 1H), 4.07 (s, 2H), 2.56 (s, 3H), 2.52 (s, 3H). C NMR (101
MHz, DMSO-D6) δ (ppm): 174.91 (s), 173.75 (s), 162.40 (s), 159.80 (s), 141.15 (s), 137.80 (s),
136.68 (d, J = 10.9 Hz), 126.59 (s), 126.00 (s), 121.16 (s), 119.10 (s), 110.81 (s), 108.09 (s), 38.48
(s), 21.01 (s), 14.21 (s). HRMS (ESI) m/z [M+ H] + calculated for C17H13BrFNO3: 378.01 found:
378.01.
(7-bromo-1-chloro-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic
acid
(2i).The
title
compound was obtained similarly to 2a. 2-(2-bromophenyl) acetic acid 7a and
2-amino-3,4-dimethylbenzoic acid 8a were replaced with 2-(2-bromo-4-chlorophenyl) acetic acid 7c and
1
2-amino-5-bromo-3,4-dimethylbenzoic acid 8b respectively. (brown solid, yield: 22.5%). H NMR
(400 MHz, DMSO-D6) δ (ppm): 13.30 (s, 1H), 9.48 (s, 1H), 8.17 (s, 1H), 7.61 (d, J = 7.9 Hz, 1H),
13
7.25 (d, J = 7.9 Hz, 1H), 4.09 (s, 2H), 2.54 (d, J = 5.9 Hz, 3H), 2.51 (s, 3H). C NMR (101 MHz,
DMSO-D6) δ (ppm): 175.55 (s), 173.44 (s), 142.01 (s), 141.25 (s), 137.57 (s), 135.83 (s), 132.73 (s),
127.02 (s), 126.03 (s), 124.68 (s), 122.82 (s), 121.44 (s), 119.15 (s), 117.12 (s), 38.57 (s), 21.05 (s),

14.19 (s). HRMS (ESI) m/z [M+ H] + calculated for C17H13BrClNO3: 395.65 found: 395.98.
(7-bromo-2-fluoro-5,6-dimethyl-9-oxo-9,10-dihydro-acridin-4-yl)-acetic
acid
(2j).The
title
compound was obtained similarly to 2a. 2-(2-bromophenyl) acetic acid 7a and
2-amino-3,4-dimethylbenzoic acid 8a were replaced with 2-(2-bromo-5-fluorophenyl) acetic acid 7d
1
and 2-amino-5-bromo-3,4-dimethylbenzoic acid 8b respectively. (yellow solid, yield: 23.4%). H
NMR (400 MHz, DMSO-D6) δ (ppm): 10.23 (s, 1H), 8.21 (s, 1H), 7.78 (dd, J = 8.8, 3.0 Hz, 1H),
7.68 (dd, J = 8.9, 3.0 Hz, 1H), 4.13 (s, 2H), 2.53 (s, 3H), 2.50 – 2.48 (m, 3H). 13C NMR (101 MHz,
DMSO-D6) δ (ppm): 178.95, 176.39, 165.29, 162.70, 145.35, 144.16, 141.50, 139.10, 135.98,
126.17, 123.32, 121.64, 113.56, 110.24, 41.10, 23.87, 15.35. HRMS (ESI) m/z [M+ H] + calculated
for C17H13BrFNO3: 378.01 found: 378.01.
2-(7-bromo-2-chloro-5,6-dimethyl-9-oxo-9,10-dihydroacridin-4-yl) acetic acid (2k). The title
compound was obtained similarly to 2a. 2-(2-bromophenyl) acetic acid 7a and
2-amino-3,4-dimethylbenzoic acid 8a were replaced with 2-(2-bromo-5-chlorophenyl) acetic acid 7e and
1
2-amino-5-bromo-3,4-dimethylbenzoic acid 8b respectively. (yellow solid, yield: 16.6%). H NMR
(400 MHz, DMSO-D6) δ (ppm): 15.19 (s, 1H), 10.59(s, 1H), 8.23 (s, 1H), 7.96 (d, J = 2.4 Hz, 1H),
7.54 (d, J = 2.4 Hz, 1H), 3.67 (s, 2H), 2.62 (s, 3H), 2.52 (s, 3H). 13C NMR (101 MHz, DMSO-D6) δ
(ppm): 176.46, 173.80, 140.98, 139.76, 138.55, 136.29, 135.21, 130.18, 129.80, 123.50, 122.08,
121.15, 120.14, 118.71, 38.96, 21.10, 14.32. HRMS (ESI) m/z [M+ H] + calculated for
C17H13BrClNO3: 395.65 found: 395.98.
2-(1-fluoro-7-methoxy-5,6-dimethyl-9-oxo-9,10-dihydroacridin-4-yl) acetic acid (2l). The
preparation of title compound was similarly to 2a. 2-(2-bromophenyl) acetic acid 7a and
2-amino-3,4-dimethylbenzoic acid 8a were replaced with 2-(2-bromo-4-fluorophenyl) acetic acid 7b
and 2-amino-5-methoxy-3,4-dimethylbenzoic acid 8c. The dark brown solid was provided in 20.8%
yield. 1H NMR (400 MHz, DMSO-D6) δ (ppm): 9.65 (s, 1H), 7.61 (dd, J = 8.1, 5.3 Hz, 1H), 7.48 (s,
1H), 6.91 (dd, J = 11.7, 8.2 Hz, 1H), 4.04 (s, 2H), 3.88 (s, 3H), 2.50 (s, 3H), 2.30 (s, 3H). 13C NMR
(101 MHz, DMSO-D6) δ (ppm): 175.51, 173.76, 162.48, 159.90, 153.56, 141.05, 135.68, 133.19,

125.21, 120.63, 118.97, 110.36, 107.22, 101.64, 55.98, 38.60, 13.32, 13.19. HRMS (ESI) m/z [M+
H] + calculated for C18H16FNO4: 330.11 found: 330.11.
2-(1-chloro-7-methoxy-5,6-dimethyl-9-oxo-9,10-dihydroacridin-4-yl) acetic acid (2m). The
preparation of title compound was similarly to 2a. 2-(2-bromophenyl) acetic acid 7a and
2-amino-3,4-dimethylbenzoic acid 8a were replaced with 2-(2-bromo-4- chlorophenyl) acetic acid 7c
and 2-amino-5-methoxy-3,4-dimethylbenzoic acid 8c. The dark brown solid was provided in 21.1%
yield. 1H NMR (400 MHz, DMSO-D6) δ (ppm): 13.15 (s, 1H), 9.41 (s, 1H), 7.57 (d, J = 7.9 Hz, 1H),
7.48 (s, 1H), 7.21 (d, J = 7.8 Hz, 1H), 4.10 (s, 2H), 3.88 (s, 3H), 2.50 – 2.46 (m, 3H), 2.31 (s, 3H).
13C NMR (101 MHz, DMSO-D6) δ (ppm): 175.99, 173.42, 153.59, 141.52, 134.93, 133.22, 133.17,
132.54, 125.11, 123.94, 122.59, 120.76, 116.46, 101.94, 55.98, 38.59, 13.31, 13.14. HRMS (ESI)
m/z [M+ H] + calculated for C18H16ClNO4: 346.08 found: 346.08.
2-(2-fluoro-7-methoxy-5,6-dimethyl-9-oxo-9,10-dihydroacridin-4-yl) acetic acid (2n). The
preparation of title compound was similarly to 2a. 2-(2-bromophenyl) acetic acid 7a and
2-amino-3,4-dimethylbenzoic acid 8a were replaced with 2-(2-bromo-5- fluorophenyl) acetic acid 7d
and 2-amino-5-methoxy-3,4-dimethylbenzoic acid 8c. The brown solid was obtained in 23.6% yield.
1H NMR (400 MHz, DMSO-D6) δ (ppm): 10.23 (s, 1H), 8.21 (s, 1H), 7.78 (dd, J = 8.8, 3.0 Hz, 1H),
7.68 (dd, J = 8.9, 3.0 Hz, 1H), 4.13 (s, 2H), 2.53 (s, 3H), 2.48 (s, 3H). 13C NMR (101 MHz,
DMSO-D6) δ (ppm): 176.12, 173.18, 155.61, 153.39, 136.18, 134.37, 133.47, 126.42, 125.31,
123.80, 121.02, 118.61, 109.28, 101.35, 55.97, 38.29, 13.40, 13.27. HRMS (ESI) m/z [M+ H] +
calculated for C18H16FNO4: 330.11 found: 330.11.
2-(2-chloro-7-methoxy-5,6-dimethyl-9-oxo-9,10-dihydroacridin-4-yl) acetic acid (2o). The
preparation of title compound was similarly to 2a. 2-(2-bromophenyl) acetic acid 7a and
2-amino-3,4-dimethylbenzoic acid 8a were replaced with 2-(2-bromo-5- chlorophenyl) acetic acid 7e
and 2-amino-5-methoxy-3,4-dimethylbenzoic acid 8c. The yellow solid was obtained in 23.1% yield.
1H NMR (400 MHz, DMSO-D6) δ (ppm): 13.21 (s, 1H), 9.51 (s, 1H), 8.11 (s, 1H), 7.76 (s, 1H), 7.50
13
(s, 1H), 4.18 (s, 2H), 3.89 (s, 3H), 2.51 (s, 3H), 2.32 (s, 3H). C NMR (101 MHz, DMSO-D6) δ

(ppm): 175.75, 173.21, 153.53, 137.93, 134.88, 134.23, 133.64, 126.14, 125.57, 125.46, 124.28,
121.30, 119.27, 101.56, 56.00, 38.11, 13.42, 13.27. HRMS (ESI) m/z [M+ H] + calculated for
C18H16ClNO4: 346.08 found: 346.08.
10-carboxymethyl-9-acridanone (CMA, 3a).
Intermediate ethyl 2-(9-oxoacridin-10(9H)-yl)
acetate was synthesized by the method of 13b. To the solution of ethyl 2-(9-oxoacridin-10(9H)-yl)
acetate (0.30g, 1.1 mmol) and NaOH (0.05 g, 1.3 mmol) in ethanol (20 ml) and water (2 ml) was
added, and the mixture was heated to 60℃ for 1h. Upon removal of the solvents, the resulting
residue was dissolved in water and filtered. After neutralization of the filtrate with conc.HCl, the
mixture was filtrated to get the crude solid, subsequently recrystallized by methanol for 0.2g white
1
power, the yield: 73.3%. H NMR (400 MHz, DMSO-D6) δ (ppm): 13.40 (s, 1H), 8.37 (d, J = 1.6
Hz, 1H), 8.35 (d, J = 1.6 Hz, 1H), 7.82 (ddd, J = 8.7, 6.9, 1.7 Hz, 2H), 7.69 (s, 1H), 7.67 (s, 1H), 7.36
(t, J = 7.4 Hz, 2H), 5.34 (s, 2H). 13C NMR (101 MHz, DMSO-D6) δ (ppm): 177.20, 170.58, 142.65
(2C), 134.77 (2C), 127.16 (2C), 122.12 (2C), 122.06(2C), 116.31 (2C), 48.02. HRMS (ESI) m/z [M+
H] + calculated for C15H11NO3: 254.07 found: 254.08.
2-methoxy-5,6-dimethyl-10-carboxymethyl-9-acridanone (3b).
The title compound was obtained
similarly to 3a. 2-(9-oxoacridin-10(9H)-yl) acetate
was
replaced
with
1
2-methoxy-5,6-dimethyl-10-(2-oxobutyl) acridin-9(10H)-one 13b. (pale red solid, yield: 78.6%). H
NMR (400 MHz, DMSO-D6) δ (ppm): 10.59 (s, 1H), 8.03 (d, J = 8.3 Hz, 1H), 7.98 (d, J = 9.2 Hz,
1H), 7.59 (d, J = 3.0 Hz, 1H), 7.41 – 7.36 (m, 1H), 7.09 (d, J = 8.3 Hz, 1H), 3.91 (s, 2H), 3.85 (s,
13
3H), 2.50 (s, 3H), 2.43 (s, 3H). C NMR (101 MHz, DMSO-D6) δ (ppm): 176.81, 174.70, 154.49,
141.53, 139.71, 136.56, 124.30, 123.90, 123.64, 123.01, 121.09, 120.55, 119.00, 105.01, 60.03,
55.81, 21.23, 13.76. HRMS (ESI) m/z [M+ H] + calculated for C18H17NO4: 312.12 found: 312.13.
2-methoxy-5,6-dimethyl-9-oxo-9,10-dihydroacridine-4-carboxylic acid (4).
The title compound
was obtained similarly to 9g. The starting materials of 7a and 8c were replaced with 10b and 8a, the
1
green product was got in the yield of 39.8%. H NMR (400 MHz, DMSO-D6) δ (ppm): 14.01 (s,
1H), 11.87 (s, 1H), 7.98 – 7.88 (m, 3H), 7.10 (d, J = 8.2 Hz, 1H), 3.89 (s, 3H), 2.42 (s, 3H), 2.36 (s,

3H). 13C NMR (101 MHz, DMSO-D6) δ (ppm): 173.82, 168.67, 148.93, 138.53, 136.73, 133.57,
121.26, 120.89, 120.65, 120.00, 118.12, 117.55, 116.03, 111.70, 59.23, 18.22, 10.75. HRMS (ESI)
m/z [M+ H] + calculated for C17H15NO4: 298.10 found: 298.11.
8-methoxy-9,10-dimethyl-6H-pyrrolo[3,2,1-de] acridine-1,6(2H)-dione (9g).
material 2-amino-5-methoxy-3,4-dimethyl-benzoic acid 8c was synthesized according to a method
previously reported in the literature.[25] To magnetically stirred solution of
The starting
a
2-(2-bromophenyl)acetic acid 7a (0.61g, 2.8mmol) and 8c (0.5g, 2.6mmol) in DMF was added
powdered Cu (0.02g), Cu2O (0.02g) and K2CO3 (0.28g, 2mmol) successively. The reaction mixture
was heated to 95℃ for 12h. Upon removal of the solvents, the residue was triturated with 1N NaOH
solution. The whole mixture was filtered off and the alkaline solution was acidified with conc. HCl,
and then collected the precipitate. The dried residue was subjected to column purification (PE/EA
4:1) to furnish the desired compound. A solution of 2-((2-(carboxymethyl) phenyl)
amino)-5-methoxy-3,4-dimethyl-benzoic acid in Eaton’s reagent (P2O5 –CH3SO3H) (5 mL) at room
temperature was heated to 90 °C for 1 h. After cooling to room temperature, the mixture was added
slowly to the saturated aqueous NaHCO3 solution. The crude was accumulated by filtration followed
by purification with silica gel column chromatography to give 0.3g product: the dark brown solid,
1
yield 37.5%. H NMR (400 MHz, CHLOROFORM-D) δ (ppm): 8.17 (d, J = 8.0 Hz, 1H), 7.81 (s,
1H), 7.62 (dd, J = 7.1, 0.9 Hz, 1H), 7.42 – 7.36 (m, 1H), 3.99 (s, 3H), 3.92 (s, 2H), 2.39 (s, 3H), 2.37
(s, 3H). 13C NMR (101 MHz, DMSO-D 6) δ (ppm): 173.51, 161.50, 155.21, 153.01, 139.05, 134.15,
132.91, 125.57, 124.99, 122.86, 121.12, 120.42, 119.14, 101.45, 55.73, 38.54, 13.16, 13.02. HRMS
(ESI) m/z [M+ H] + calculated for C18H15NO3: 294.11 found: 294.11.
General Procedure for the Preparation of acridin-9(10H)-one (12a).
To a solution of
2-bromo-benzoic acid 10a (1.29g, 6.4mmol) and aniline 11a (0.6g, 6.4mmol) in DMF (6ml) at room
temperature, we subsequently added powdered Cu (0.05g), Cu2O (0.05g) and K2CO3 (0.71g,
5.1mmol) was added. The reaction mixture was heated at 110℃ for 12h. Upon removal of the
solvents under a vacuum, the residue was dissolved in 1N NaOH solution(25ml). The crude product
was obtained by precipitation upon acidification of the filtrate with conc.HCl. After drying, the crude

2-(phenylamino) benzoic acid was added to Eaton’s reagent (5 mL) at room temperature, then the
mixture was heated to 90 °C for 1 h. The cooled reaction mixture was dropped into the saturated
aqueous NaHCO3 solution. The precipitate was filtered to collect the rough product. Purification of
the residue by silica gel column chromatography provided compound 12a, an off-white solid. (0.47
g, 37.6% yield). 1H-NMR (400 MHz, DMSO-D 6) δ (ppm): 11.75 (s, 1H), 8.28 – 8.20 (m, 2H), 7.74
13
(ddd, J = 8.4, 6.9, 1.5 Hz, 2H), 7.55 (d, J = 8.3 Hz, 2H), 7.32 – 7.23 (m, 2H). C NMR (101 MHz,
DMSO-D 6) δ (ppm): 177.30, 141.41 (2C), 133.99 (2C), 126.54 (2C), 121.53 (2C), 121.01 (2C),
117.87 (2C). HRMS (ESI) m/z [M+ H] + calculated for C13H9NO: 196.07 found: 196.08.
2-methoxy-5,6-dimethylacridin-9(10H)-one (12b).
The preparation of title compound was
similarly to 12a. The compound 12b was obtained from 2-bromo-5-methoxybenzoic acid 10b and
2,3-dimethylaniline 11b, as a pale yellow solid in 35.5% yield. 1H-NMR (400 MHz, DMSO-D 6 ) δ
(ppm): 10.44 (s, 1H), 8.03 (d, J = 8.3 Hz, 1H), 7.90 (d, J = 9.1 Hz, 1H), 7.60 (d, J = 3.0 Hz, 1H), 7.38
(dd, J = 9.1, 3.0 Hz, 1H), 7.08 (d, J = 8.3 Hz, 1H), 3.86 (s, 3H), 2.47 (s, 3H), 2.41 (s, 3H). 13C NMR
(101 MHz, DMSO-D6) δ: 176.82, 154.51, 141.53, 139.73, 136.57, 124.26, 123.89, 123.65, 123.00,
121.12, 120.55, 119.03, 105.10, 55.84, 21.22, 13.75. HRMS (ESI) m/z [M+ H] + calculated for
C16H15NO2: 254.11 found: 254.12.
2-methoxyacridin-9(10H)-one (12c).
The preparation of title compound was similarly to 12a.
The compound 12c was obtained from 2-bromo-5-methoxybenzoic acid 10b and aniline 11a, as a
pale green solid in 40.4% yield. 1H-NMR (400 MHz, DMSO-D 6 ) δ (ppm): 11.74 (s, 1H), 8.23 (dd,
J = 8.2, 1.4 Hz, 1H), 7.71 (ddd, J = 8.4, 6.9, 1.5 Hz, 1H), 7.63 (d, J = 2.9 Hz, 1H), 7.54 (d, J = 3.2
Hz, 1H), 7.52 (d, J = 2.4 Hz, 1H), 7.42 (dd, J = 9.0, 3.0 Hz, 1H), 7.24 (ddd, J = 8.0, 6.9, 1.0 Hz, 1H),
13
3.86 (s, 3H). C NMR (101 MHz, DMSO-D 6) δ (ppm): 176.64, 154.47, 140.97, 136.24, 133.55,
126.47, 124.82, 121.50, 121.23, 120.15, 119.72, 117.83, 105.37, 55.87. HRMS (ESI) m/z [M+ H] +
calculated for C14H11NO2: 226.08 found: 226.09.
3,4-dimethylacridin-9(10H)-one (12d).
The preparation of title compound was similarly to 12a.
The compound 12d was obtained from 2-bromo-benzoic acid 10a and 2,3-dimethylaniline 11b, as a

1
yellow-green solid in 45.3% yield. H-NMR (400 MHz, DMSO-D 6 ) δ (ppm): 10.47 (s, 1H), 8.20
(dd, J = 8.1, 1.3 Hz, 1H), 8.03 (d, J = 8.2 Hz, 1H), 7.93 (d, J = 8.4 Hz, 1H), 7.73 – 7.67 (m, 1H), 7.25
13
(ddd, J = 8.0, 7.0, 0.9 Hz, 1H), 7.11 (d, J = 8.2 Hz, 1H), 2.49 (s, 3H), 2.42 (s, 3H). C NMR (101
MHz, DMSO-D6) δ (ppm): 173.82, 138.53, 137.60, 136.73, 127.67, 121.26, 120.89, 120.65, 120.00,
119.44, 118.12, 117.55, 116.03, 18.22, 10.75. HRMS (ESI) m/z [M+ H] + calculated for C15H13NO:
224.10 found: 224.11.
2-methoxy-3,4-dimethylacridin-9(10H)-one(12e).
The preparation of title compound was similarly
to 12a. The compound 12d was obtained from 2-bromo-benzoic acid 10a and
1
4-methoxy-2,3-dimethylaniline 11c, as a dark green solid in 38.8% yield. H-NMR (400 MHz,
DMSO-D 6 ) δ (ppm): 10.46 (s, 1H), 8.20 (d, J = 7.9 Hz, 1H), 7.90 (d, J = 8.4 Hz, 1H), 7.68 (t, J =
7.6 Hz, 1H), 7.55 (s, 1H), 7.22 (t, J = 7.2 Hz, 1H), 3.88 (s, 3H), 2.52 (s, 3H), 2.30 (s, 3H). 13C NMR
(101 MHz, DMSO-D 6) δ (ppm): 176.66, 152.83, 141.21, 134.90, 133.01, 132.79, 126.01, 125.24,
121.25, 119.91, 119.60, 118.64, 101.72, 55.90, 14.27, 13.35. HRMS (ESI) m/z [M+ H] + calculated
for C16H15NO2: 254.11 found: 254.12.
2,7-dimethoxy-3,4-dimethylacridin-9(10H)-one(12f).
The preparation of title compound was
similarly to 12a. The compound 12d was obtained from 2-bromo-5-methoxybenzoic acid 10b and
1
4-methoxy-2,3-dimethylaniline 11c, as a black-green solid in 31.7% yield. H-NMR (400 MHz,
DMSO-D 6 ) δ (ppm): 10.48 (s, 1H), 7.89 (d, J = 9.1 Hz, 1H), 7.58 (d, J = 2.9 Hz, 1H), 7.53 (s, 1H),
13
7.36 (dd, J = 9.1, 3.0 Hz, 1H), 3.87 (s, 3H), 3.85 (s, 3H), 2.50 (s, 3H), 2.30 (s, 3H). C NMR (101
MHz, DMSO-D6) δ (ppm): 175.84, 154.26, 152.69, 136.11, 134.63, 132.46, 125.10, 124.02, 120.53,
120.30, 118.73, 104.50, 101.37, 55.89, 55.76, 14.24, 13.33. HRMS (ESI) m/z [M+ H] + calculated
for C17H17NO3: 284.12 found: 284.13.
2-methoxy-5,6-dimethyl-10-(2-oxobutyl) acridin-9(10H)-one(13b).
To a solution of NaH
(1.43mmol) and compound 12b (0.33g, 1.3mmol) in DMF (5ml) at room temperature was stirred for
1 h and then cooled to 5–7℃. The ethyl bromoacetate (0.43ml, 2.6mmol) was added to the resulting
mixture and continuously stirred at room temperature for 20 h. After completion of the reaction

(TLC), the reaction mixture was poured into ice water (15ml). The resulting precipitates were filtered
off, dried, and then extracted with chloroform. Evaporation of the solvent gave crude ester which
1
was purified by recrystallization. As a green solid (0.33g) in yield of 76%. H NMR (400 MHz,
DMSO-D6) δ 8.12 (d, J = 8.8 Hz, 1H), 8.06 (d, J = 9.4 Hz, 1H), 7.63 (d, J = 2.8 Hz, 1H), 7.50–7.43
(m, 2H), 5.05 (s, 2H), 4.25 (q, J = 7.1 Hz, 2H), 3.97 (s, 3H), 2.77 (s, 3H), 2.51 (s, 3H), 1.26 (d, J =
13
7.1 Hz, 3H). C NMR (101 MHz, DMSO-D 6) δ (ppm): 169.40, 158.26, 156.85, 147.94, 146.48,
136.71, 133.75, 131.85, 129.78, 125.47, 119.97, 119.40, 118.75, 98.58, 72.24, 61.45, 56.04, 20.84,
14.58, 13.90. HRMS (ESI) m/z [M+ H] + calculated for C20H21NO4: 340.15 found: 340.16.
4.2. Biological Evaluations
4.2.1. In vitro cell-based Assay
The following reporter cell lines and positive compound were obtained from InvivoGen:
293T-hSTING-R232 Cells, 293T-mSTING Cells, THP1-KO-STING and 2′,3′-cGAMP. The setting
of concentration gradient as follows: the targeted small molecule was prepared for ten different
concentrations (3.5 μM, 5.2 μM, 7.8 μM, 11.7 μM, 17.6 μM, 26.3 μM, 39.5 μM, 59.2 μM, 88.8 μM,
133.3 μM, and 200 μM), and 2′,3′-cGAMP was set into eight different concentrations (0.27 μM, 0.54
μM, 1.08 μM, 2.16 μM, 4.32 μM, 8.64 μM, 17.3 μM, and 34.6 μM). The general procedure for
screening of targeted compounds for cytokine induction activity: Firstly, 180μL of 293T cells
(∼50000 cells per well) or THP-1 cells (∼100000 cells per well) were stimulated with 20μL of either
saline or a saline solution of a test compound for 48 h at 37 °C in a 5% CO2 incubator. Then, the
inductions of cytokines were indirectly quantified using QUANTI-Blue or QUANTI-Luc, which
were prepared and used according to the manufacturer’s instructions. Finally, EC50 values were
calculated using software Origin 8.0, and Emax values were normalized to the response induced by
2′,3′-cGAMP in each assay. The results are expressed as the mean ± SD from three independent
experiments.
4.2.2. HTRF ® Human STING binding competitive Assay
By using the human STING binding kits purchased from Cisbio, the human STING binding